Schizophrenia Research xxx (2014) xxx–xxx

SCHRES-05761; No of Pages 7

Contents lists available at ScienceDirect

Schizophrenia Research

j ourna l homepage: www.e lsev ie r .com/ locate /schres

Emotion recognition as a predictor of transition to a psychotic disorder inultra-high risk participants

Kelly A. Allott a,⁎, Miriam R. Schäfer a,b, Andrew Thompson a,c, Barnaby Nelson a, Sarah Bendall a,Cali F. Bartholomeusz d, Hok Pan Yuen a, Patrick D. McGorry a, Monika Schlögelhofer b,Andreas Bechdolf a,e, G. Paul Amminger a,b

a Orygen Youth Health Research Centre, Centre for Youth Mental Health, The University of Melbourne, Australiab Department of Child and Adolescent Psychiatry, Medical University Vienna, Austriac East Sussex Early Intervention in Psychosis Service, Sussex Partnership NHS Foundation Trust, United Kingdomd Melbourne Neuropsychiatry Centre, The University of Melbourne, Australiae Department of Psychiatry and Psychotherapy, University of Cologne, Germany

⁎ Corresponding author at: Orygen Youth Health ResParkville, Victoria 3052, Australia. Tel.: +61 3 9342 2942;

E-mail address: kallott@unimelb.edu.au (K.A. Allott).

0920-9964/$ – see front matter © 2014 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.schres.2014.01.037

Please cite this article as: Allott, K.A., et al., EmSchizophr. Res. (2014), http://dx.doi.org/10.

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 19 June 2013Received in revised form 11 December 2013Accepted 14 January 2014Available online xxxx

Keywords:Affect recognitionNeutral emotionFearTrait markerEarly psychosisConversion

Aims: Recent research has shown emotion recognition to be impaired in individuals at ultra-high risk (UHR)for developing a psychotic disorder compared to healthy controls. This longitudinal study aimed to examinewhether disturbed emotion recognition measured in UHR participants at baseline predicts transition to a psy-chotic disorder within 12 months.Methods: Thirty-seven UHR participants aged 13–22 years participated in the study. At baseline participantscompleted face and prosody emotion recognition tasks, as well as measures of psychopathology, functioning,and IQ. Transition to a psychotic disorder over 12 months was the primary outcome. A series of Cox regressionswas performedwith emotion recognition as the predictor variable, while controlling for covariates, with time totransition to a psychotic disorder as the dependent variable.Results: Eleven (29.7%) of the 37 participants transitioned to a psychotic disorder over the 12-month follow-upperiod. Total face or prosody emotion recognition accuracy was not predictive of transition to a psychotic disor-der. However, examination of recognition of specific emotions, while controlling for positive, negative and global

symptoms and functioning, revealed that accuracy in identifying neutral (p= .037) and fearful (p= .015) emo-tion predicted transition to a psychotic disorder. Specifically, lower accuracy in identifying neutral emotion andhigher accuracy in identifying fearful emotion were predictive of transition to a psychotic disorder within12 months. Examination of the separate modalities revealed that this finding held for face but not for prosodyemotion recognition.Conclusion: These findings suggest that emotion recognition abilities may be prognostic for the development ofpsychotic disorders, but further studies are needed.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The neurodevelopmental hypothesis proposes that the etiologyof schizophrenia involves abnormal brain development resulting fromgenetic and environmental influences (Marenco and Weinberger,2000). Central to this theory is the notion that neurological, cognitiveand behavioral deficits are likely to emerge during childhood and/oradolescence and precede full-threshold psychotic disorder. Accordingly,research in individuals deemed to be at clinical high-risk (CHR) for apsychotic disorder aims to identify neurobehavioral markers that maybe predictive of psychosis onset, provide insight into the pathogenesisof psychosis, and offer potential targets for preventive treatments.

earch Centre, 35 Poplar Road,fax: +61 3 9342 2858.

ghts reserved.

otion recognition as a predic1016/j.schres.2014.01.037

Regarding cognition, most studies to date have examined neurocognition(Brewer et al., 2006). However, social cognition, a related but indepen-dent construct, has received markedly increased attention in recentyears (Green and Phillips, 2004; Thompson et al., 2011). In particular,the domain of emotion recognition has been extensively studied(Couture et al., 2006; Kohler et al., 2010).

Interpersonal encounters involve the perception, decoding and re-sponse to emotive cues (Green and Phillips, 2004). One's ability to accu-rately decipher and integrate these cues with other relevant socialinformation influences social competence (Couture et al., 2006). Stableand significant deficits in facial and prosody emotion recognition havebeen demonstrated in first-episode and chronic schizophrenia samples(Edwards et al., 2002; Kucharska-Pietura et al., 2005; Kohler et al.,2010; Green et al., 2012) and are associated with poorer social function-ing (Irani et al., 2012). Although longitudinal studies are limited, thesefindings tend to suggest that emotion recognition abnormalities might

tor of transition to a psychotic disorder in ultra-high risk participants,

2 K.A. Allott et al. / Schizophrenia Research xxx (2014) xxx–xxx

represent a trait marker (as opposed to state marker/symptom) ofpsychotic disorders.

Supporting evidence for this premise comes from studies demon-strating that emotion recognition deficits are also present in CHR orultra-high risk (UHR) for psychosis populations (Addington et al.,2008; van Rijn et al., 2011; Amminger et al., 2012a,c; Green et al.,2012). On a facial emotion recognition task, Addington et al. (2008)found that CHR individuals performed significantly worse than healthycontrols and their performance was not significantly different fromfirst- and multiple-episode psychosis patients. Green et al. (2012) alsofound poorer emotion identification in CHR individuals versus healthycontrols, but their performance was significantly better than recent-onset and chronic psychosis patients. In adolescent UHR individuals,van Rijn et al. (2011) found impaired labeling of neutral faces relativeto healthy controls, with a tendency for UHR participants to misattrib-ute neutral faces as angry. We recently examined emotion recognitionfor faces and voices in UHR individuals and found impairments acrossboth sensory modalities (Amminger et al., 2012b). Specifically, theability to recognize fear and sadness (modalities combined) was signif-icantly poorer in UHR individuals than healthy controls, and these defi-cits were of similar severity as those with first-episode schizophrenia(Amminger et al., 2012b). Specific examination of facial emotion recog-nition revealed the same pattern of results; however, in voices onlyrecognition of anger was impaired in both clinical groups.

Several studies have also reported significant emotion recogni-tion deficits in unaffected first-degree relatives of individuals withschizophrenia-spectrum disorders (Kee et al., 2004; Bediou et al.,2007; Eack et al., 2010). Although Eack et al. (2010) found no significantdifference in total emotion recognition scores between unaffected fam-ilymembers andhealthy controls, analysis of specific emotions revealeda significant deficit in labeling neutral faces in the unaffected familymembers, with a tendency to label neutral emotions negatively.

Collectively, results of these studies suggest that emotion recogni-tionmay represent a trait marker of psychotic disorders. However, find-ings of cross-sectional high-risk studies are diluted by the fact that mostindividuals in these cohorts do not go on to develop a psychotic disorder(Yung et al., 2007) and emotion recognition abnormalities are observedin other psychiatric disorders (Derntl et al., 2009; Demenescu et al.,2010; Kohler et al., 2011). Thus, it is possible that emotion recognitionabnormalities are simply epiphenomena or trait markers of generalpsychopathology. Stronger evidence would come from longitudinalprospective studies that examine whether emotion recognition is pre-dictive of transition to a psychotic disorder. Pinkham et al. (2007) re-ported similar facial emotion recognition performance in a sample ofat-risk individuals who transitioned to a psychotic disorder (n = 5)within 12 months versus those who did not (n = 14), suggesting thatemotion recognition may not be predictive of transition. However,significance testing was not performed due to the small sample.Addington et al. (2012) examined face and prosody emotion recogni-tion in a large longitudinal study of CHR individuals. They found no dif-ference between the baseline emotion recognition scores of participantswho transitioned (n = 25) and those who did not transition to apsychotic disorder over 2 years (n = 121); but more than half (65%)of the original samplewas not followed-up (for various reasons), signif-icantly limiting the conclusions that could be drawn. Furthermore, re-sults of the predictive value of individual emotions were not reportedand the effect of time to follow-upwas not accounted for in the analysis.

The current longitudinal study aimed to examine whether baselineemotion recognition predicts transition to a psychotic disorder over12 months in UHR individuals. In accordance with findings in schizo-phrenia and first-episode psychosis, we first hypothesized that poorertotal face and total prosody emotion recognition would be predictiveof transition to a psychotic disorder. Second, based on our previousfind-ings (Amminger et al., 2012b), we predicted that deficits in recognizingthe specific emotions of anger, fear and sadness would be predictive oftransition. Third, in accordancewith other recent literature highlighting

Please cite this article as: Allott, K.A., et al., Emotion recognition as a predicSchizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.01.037

the significance of neutral emotion recognition (Eack et al., 2010; vanRijn et al., 2011), we hypothesized that misattribution of neutral emo-tion would be predictive of transition to a psychotic disorder.

2. Methods

2.1. Sample

The sample was derived from the 40 UHR participants assignedto the placebo treatment arm of a randomized controlled trial thatexamined whether long-chain omega-3 polyunsaturated fatty acidscould prevent the onset of psychotic disorder, described elsewhere(Amminger et al., 2010). Similar to a previous analysis concerned withthe prediction of psychosis from this trial (Amminger et al., 2012b),only theparticipantswho received placebowere included in thepresentinvestigation because the transition rates differed significantly betweenthe two treatment groups (Amminger et al., 2010). One additionalparticipant assigned to the omega-3 group who mistakenly receivedplacebo was also included in the current study. Of the 41 includedparticipants, two had no follow-up data and two did not complete theemotion recognition measures. Thus, complete data at baseline andfollow-up were available for 37 (90.2%) participants.

The study was conducted at the psychosis detection and treatmentunit of the Department of Child and Adolescent Psychiatry, MedicalUniversity Vienna (May 2004–May 2007). Participants were eligible ifthey were aged 13–25 years and met criteria for one or more of threeoperationally-defined UHR groups (Yung et al., 1998): attenuated posi-tive psychotic symptoms (group 1); transient psychosis (group 2); orgenetic risk plus decreased functioning (group 3). The presence ofattenuated or transient psychotic symptoms was determined with thePositive and Negative Syndrome Scale (PANSS; Kay et al., 1987) usingvalidated symptom severity, frequency and duration criteria (Yunget al., 1998; Morrison et al., 2004). The third group comprised individ-uals with a first-degree relative with a history of psychotic disorder(Andreasen et al., 1977) or having a DSM-IV schizotypal personality dis-order and a ≥30% decrease in functioning on the Global Assessment ofFunctioning (GAF) scale. Participants were assessed by two psychia-trists within 1 week of initial presentation and were antipsychotic-naïve at baseline and throughout the 12-month follow-up unless theydeveloped a psychotic disorder during this time.

Exclusion criteria were: impaired vision (blurred or b20/20 withcorrection); impaired auditory acuity; organic mental disorder; IQb 70; or neurological disorder. The study was approved by the EthicsCommittee of theMedical University of Vienna. All participants providedwritten informed consent, including parental consent for thoseb18 years of age.

2.2. Measures

2.2.1. Transition to a psychotic disorderThe primary outcome was transition to a psychotic disorder within

12 months, defined as a PANSS score of ≥4 on hallucinations or delu-sions, or ≥5 on conceptual disorganization, maintained for ≥1 week(Yung et al., 1998; Morrison et al., 2004). Transition was independentlyconfirmed by psychiatrists not associated with the study. TheStructured Clinical Interview for DSM-IV-TR Axis I Disorders (Firstet al., 2001) was used to determine psychiatric diagnoses, includingmajor depressive disorder (MDD) and anxiety disorders. Further detailsof the follow-up process and monitoring for transition is describedelsewhere (Amminger et al., 2010).

2.2.2. Emotion recognitionFace and voice (prosody) emotion recognition was examined at

baseline. The facial emotion recognition task was a computerizedmodification of Feinberg et al.'s (1986) procedure. Stimuli consisted of21 slides of photographs of faces, each posing one of seven emotions:

tor of transition to a psychotic disorder in ultra-high risk participants,

Table 1Baseline demographics and clinical characteristics of UHR individuals who transitioned toa psychotic disorder and those who did not over 12 months.

Transition(n = 11)

Non-transition(n = 26)

p valuea

DemographicsSex (% male) 36.4% 30.8% .513Age (mean, SD) 15.9 (1.1) 16.3 (1.9) .500Education (%)

9 years 54.5% 30.8% .161≥11 years 45.5% 69.2%

IQ (mean, SD) 106.5 (17.9) 97.3 (14.4) .109

Clinical variablesUHR group (%)Group 1 63.6% 53.8% .610Group 2 0% 11.5%Groups 1 + 2 27.3% 30.8%Groups 1 + 3 9.1% 3.9%

DUP (months; mean, SD) 10.5 (11.1) 13.4 (12.4) .496MDD (% yes) 54.5% 38.5% .367Anxiety disorder (% yes) 54.5% 30.8% .173BPD (% yes) 27.3% 15.4% .399Any substance use (% yes) 36.4% 15.4% .157MADRS (mean, SD) 22.3 (7.3) 18.0 (9.2) .179PANSS— Positive(mean, SD)

15.5 (2.0) 13.6 (3.2) .084

PANSS— Negative(mean, SD)

17.1 (7.7) 12.3 (5.4) .037

PANSS— Global(mean, SD)

34.5 (6.8) 27.8 (5.8) .004

GAF (mean, SD) 50.9 (11.1) 62.1 (12.2) .013

SD = standard deviation; DUP = duration of untreated psychosis (months); UHR =ultra-high risk; Group 1= attenuated positive psychotic symptoms; Group 2= transientpsychosis; Group 3 = genetic risk plus a decrease in functioning; MDD= Major Depres-sive Disorder; BPD = Borderline Personality Disorder; MADRS = Montgomery-AsbergDepression Rating Scale; PANSS = Positive and Negative Syndrome Scale; GAF = GlobalAssessment of Functioning.

a T-test for continuous and chi-square for categorical data.

3K.A. Allott et al. / Schizophrenia Research xxx (2014) xxx–xxx

sadness, anger, happiness, disgust, surprise, fear, and neutrality; eachexposed for 0.5 s. Prosody stimuli were used from a task developed byEdwards et al. (2001). Three actors spoke simple sentences in differentemotional tones, comprising 60 items across five emotions: fear, sad-ness, anger, surprise, and neutrality. Eight-second pauses occurredbetween items. Each speaker and task was preceded by three practiceitems. A forced-choice response format was used in both emotion rec-ognition tasks. Scores were calculated as percentage correct. Furtherdetails are provided elsewhere (Amminger et al., 2012c).

2.2.3. Clinical measuresThe PANSS (Kay et al., 1987) was used to rate negative, positive and

global symptoms, the Montgomery-Asberg Depression Rating Scale(MADRS) (Montgomery and Asberg, 1979)measured depressive symp-toms, and the GAF measured functioning. Raters were experiencedclinicians who were well trained in administration of all measures.Inter-rater reliability estimates were excellent (intra-class correlationcoefficients 0.92–0.99). Additionally, IQ was estimated with theNumber-Combination Test (Oswald and Roth, 1987). Diagnosis ofborderline personality disorder (BPD) was determined via consensusof two psychiatrists and a senior clinical psychologist through strict ap-plication of DSM-IV criteria using information from all available infor-mation sources, including medical records and discharge summaries,interviews with relatives, and longitudinal information from treatmentsessions.

2.3. Statistical analyses

Concordantwith research showing a relationship between facial andprosody emotion recognition (Kucharska-Pietura et al., 2005), we calcu-lated summary scores across the emotion modalities, taking intoaccount the different number of face and prosody tasks. However,given research indicating differences in the effectiveness with whichthe face and voice convey different emotions (Edwards et al., 2002)(e.g., happiness is the easiest facial emotion to recognize; anger is easierto recognize in voice than face), subsequent analyses were also per-formed on the separate emotion modalities.

Independent samples t-test and chi-square analyses comparedparticipants who had developed a psychotic disorder versus those whohad not on baseline demographic and clinical variables. Unadjusted sur-vival analysis using univariate Cox regression was used to examine therelationship between emotion recognition scores and rate of transitionto a psychotic disorder. Significant emotion recognition predictors(p ≤ .05) were subsequently re-entered into a Cox regression, control-ling for baseline demographic and clinical variables that differedbetween groups at a conservatively a priori specified probability valueof p ≤ .10 (see Results). A conservative p value of .10 was chosen toensure all potentially relevant variables (covariates) were included inthe Cox regression models.

For ease of interpretation, baselinemeans and standard deviations ofthe emotion recognition scores in thosewho transitioned to a psychoticdisorder and those who did not were calculated, including effect sizesusing Cohen's d methodology (1988). Statistical tests were 2-tailed.Analyses were performed using IBM SPSS, version 20.0.

3. Results

3.1. Sample characteristics

The mean age of the sample was 16.2 years (SD = 1.7; range13–22 years). Twenty-five (67.6%) of the participants were female.Regarding education, 14 (37.8%) had completed nine years, 22 (59.5%)had completed 11–13 years and 1 (2.7%) had completed tertiary-leveleducation. The mean IQ of the sample was 100.1 (SD = 15.9; range70–130). The UHR criterion represented by the sample was as follows:21 (56.8%) attenuated psychotic symptoms (group 1); 3 (8.1%)

Please cite this article as: Allott, K.A., et al., Emotion recognition as a predicSchizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.01.037

transient psychosis (group 2); 0 genetic risk plus decreased functioning(group 3); 11 (29.7%) attenuated psychotic symptoms plus transientpsychosis (groups 1 & 2); and 2 (5.4%) attenuated psychotic symptomsplus genetic risk plus decreased functioning (groups 1 & 3).

3.2. Transition to a psychotic disorder

Of the 37 participants, 11 (29.7%) transitioned to a psychotic disor-der over 12 months. The psychotic diagnoses of those who transitionedwere: 8 (73%) schizophrenia, 1 (9%) schizophreniform disorder, 1 (9%)schizoaffective disorder, and 1 (9%) bipolar I disorder. Table 1 showsbaseline demographics and clinical characteristics comparing thosewho transitionedwith those who did not. Participantswho transitionedhad significantly higher positive, negative and global PANSS symptomsand a lowerGAF score at baseline (p b .10). These variableswere includ-ed as covariates in the final prediction model.

3.3. Prediction of transition to a psychotic disorder

Using Cox regression, neither total face nor total prosody emotionrecognition (emotions combined) was predictive of transition to apsychotic disorder (p = .948 and p = .795, respectively). UnivariateCox regressions with the specific emotions (combined across facesand prosody) entered as predictor variables revealed that accuracy inidentifying neutral (p= .037) and fearful (p= .049) emotion predictedtransition to a psychotic disorder. Specifically, UHR individualswhowereless accurate at identifying neutral emotion and those who were moreaccurate at identifying fearful emotion were more likely to transition.In the final multivariate model, neutral and fearful emotion recognition(modalities combined) remained predictive of transition to a psychoticdisorder after controlling for baseline positive, negative and global

tor of transition to a psychotic disorder in ultra-high risk participants,

Table 3Cox regression model showing recognition of facial neutral and fearful emotion aspredictive of transition to a psychotic disorder, controlling for positive, negative andglobal symptoms and functioning.

Variable Beta SE Wald p value

PANSS— Positive .276 .157 3.119 .077PANSS— Negative −.042 .079 0.362 .547PANSS— Global .141 .103 1.894 .169GAF −.082 .047 3.082 .079Neutral emotion recognitiona −7.481 2.658 7.921 .005Fear emotion recognitiona 3.184 1.552 4.206 .040

PANSS = Positive and Negative Syndrome Scale; GAF = Global Assessment ofFunctioning.

a Faces emotion recognition.

4 K.A. Allott et al. / Schizophrenia Research xxx (2014) xxx–xxx

symptoms and functioning (Table 2). To determinewhichmodality (faceor prosody recognition) might account for this finding we subsequentlyexamined fear and neutral emotion recognition in faces and voices sep-arately (with adjustment for positive, negative and global symptomsand functioning). Consistent with the combined modality score, facialneutral and fear emotion recognition predicted transition to a psychoticdisorder (neutral p = .005; fear p = .040; Table 3). However, prosodicemotion recognition for both emotions was not predictive (neutralp = .547; fear p = .288).

Given that a neutral emotion recognition deficitwas found to be pre-dictive of transition, we examined whether this deficit was associatedwith a bias towardsmislabeling neutral emotion (modalities combined)as a specific emotion (Table 4). Logistic regression was used on eachemotion to test for the difference between those who did and did nottransition to a psychotic disorder in incorrectly labeling neutral emo-tion. Therewas a significant difference between the two groups formis-attribution of fear, with the group who transitioned showing a highertendency tomislabel neutral emotion as fear (8.5% vs.1.8%, respectively;p b .001). We performed additional analyses to test if these findingswere due to an increased tendency or bias in labeling all emotions asfearful and found that this was not supported (results available onrequest).

For ease of interpretability, Table 5 shows mean percentages ofcorrect emotion labeling in those who did and did not transition to apsychotic disorder. The effect sizes for differences between the twogroups in neutral and fear emotion recognition (face and prosody com-bined) were medium to large (Cohen, 1988).

3.4. Supplementary analyses

Although the groups did not differ in the frequency of comorbidpsychiatric disorders (specifically, MDD, anxiety disorder and BPD;Table 1), given that emotion recognition abnormalities may also bepresent in non-psychotic disorders, we re-ran separate Cox regressionswith these comorbid diagnoses added as covariates. The main findingsremained unchanged. Specifically, when MDD was included, both neu-tral (p= .029) and fear (p= .041) emotion recognition remained pre-dictive of transition. With anxiety disorder included as a covariate,again, recognition of both emotions remained significant (neutralp = .019; fear p = .034). Similarly, the inclusion of BPD diagnosis didnot alter the findings (neutral p = .025; fear p = .037). Additionally,since both IQ and age may influence emotion recognition ability; thesevariables were also subsequently included separately as covariates inthe regression model. Interestingly, when IQ was included, it wasfound to be a significant predictor of transition (p= .032), with higherIQ being predictive. Nevertheless, while controlling for IQ, both neutral(p = .014) and fear (p = .027) emotion recognition remained predic-tive of transition. Age was not a significant predictor (p = .079), butboth emotions remained significant predictors of psychosis transition(neutral p = .017; fear p = .025) while controlling for age.

Table 2Final Cox regression model showing recognition of neutral and fearful emotion aspredictors of transition to a psychotic disorder, controlling for positive, negative andglobal symptoms and functioning.

Variable Beta SE Wald p value

PANSS— Positive .291 .157 3.422 .064PANSS— Negative .042 .086 0.237 .626PANSS— Global −.021 .127 0.027 .869GAF −.076 .047 2.572 .109Neutral emotion recognitiona −.113 .050 4.337 .024Fear emotion recognitiona .080 .038 4.337 .037

PANSS = Positive and Negative Syndrome Scale; GAF = Global Assessment ofFunctioning.

a Combined modality score.

Please cite this article as: Allott, K.A., et al., Emotion recognition as a predicSchizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.01.037

4. Discussion

This study examinedwhether emotion recognitionwas predictive oftransition to a psychotic disorder over 12 months in antipsychotic-naïve UHR individuals. Contrary to our first hypothesis, total face andprosody emotion recognition performance (combining emotions) wasnot predictive of transition. Furthermore, opposing our second hypoth-esis, deficits in recognizing negative emotions (anger, fear, sadness)were also not predictive of transition. Interestingly, although we previ-ously found UHR individuals as a group to have deficits in recognizinganger, fear and sadness as compared to healthy controls (Ammingeret al., 2012c), the current study found that better fear identificationwas predictive of transition. Our final hypothesis that poorer identifica-tion of neutral emotion would be predictive of transition to a psychoticdisorder was supported. Both fear and neutral emotion recognitionremained significant predictors after controlling for positive, negativeand global symptoms and functioning. Subsequent analyses of theseparate emotion modalities revealed that the finding regarding fearand neutral emotions remained for face, but not prosody emotionrecognition. To our knowledge, this is the first study to show emotionrecognition to be predictive of transition to a psychotic disorder.

Total face and voice emotion recognition not predicting transition toa psychotic disorder is consistent with Addington et al.'s (2012) study.Although global emotion recognition deficits have been reportedin UHR samples (Addington et al., 2008; Green et al., 2012), the currentfindings highlight the value of examining accuracy of labeling specificemotions, as shown in previous studies (Kohler et al., 2003; Ammingeret al., 2012c). However, in our previous study, althoughwe found signif-icant deficits in recognizing negative emotions in UHR individuals com-pared to healthy controls (Amminger et al., 2012c), these deficits alonedid not predict which UHR individuals would later develop a psychoticdisorder in the current study. It is possible that impaired negative emo-tion recognition may be a non-specific deficit associated with variouspsychopathologies and is not necessarily specific to psychotic disorders.Indeed, deficits or alterations in emotion recognition have been found inmood and anxiety disorders (Derntl et al., 2009; Demenescu et al., 2010;Kohler et al., 2011). It may be that the pattern of emotion recognition

Table 4Mean percentages and standard deviations for errors in labeling neutral emotion(modalities combined) in those who did and did not transition to a psychotic disorder.

Transition (n = 11) Non-transition (n = 26) p valuea

Neutral emotion labeled asSad 16.4 (5.5) 17.2 (12.8) 0.814Anger 1.2 (2.7) 2.1 (3.1) 0.481Fear 8.5 (9.5) 1.8 (4.4) b0.001Disgust 6.1 (13.5) 1.3 (6.5) 0.181Surprise 1.8 (4.3) 0.3 (1.3) 0.061

Note: No participant mislabelled neutral emotion as happy. All emotions are combinedacross modalities, except disgust, which is for faces only.

a The p-values were obtained by applying logistic regression on each emotion.

tor of transition to a psychotic disorder in ultra-high risk participants,

Table 5Mean percentages (and standard deviations) of correct emotion recognition at baseline inUHR individuals who transitioned to a psychotic disorder and those who did not.

Transition (n = 11) Non-transition (n = 26) Effect size (d)a

Combined emotion recognitionTotal 74.2 (8.5) 73.8 (8.2) 0.05Sadness 61.4 (20.2) 60.4 (14.3) 0.06Fear 72.3 (14.8) 60.4 (17.2) 0.74Anger 78.0 (13.8) 82.5 (14.0) −0.32Surprise 83.0 (16.6) 81.9 (17.0) 0.07Neutral 76.1 (10.9) 83.8 (9.9) −0.74

Facial emotion recognitionTotal 77.1 (8.7) 77.1 (10.3) 0.00Sadness 69.7 (23.4) 69.2 (16.12) 0.03Fear 81.8 (22.9) 59.0 (33.1) 0.81Anger 60.6 (25.0) 69.2 (24.8) −0.36Surprise 81.8 (22.9) 85.9 (23.4) −0.18Neutral 84.9 (22.9) 93.6 (13.4) −0.48Happiness 97.0 (10.1) 100.0 (0.0) −0.30Disgust 63.6 (31.5) 62.8 (25.5) 0.03

Prosody emotion recognitionTotal 72.6 (10.4) 72.2 (7.9) 0.04Sadness 53.0 (26.7) 51.6 (19.6) 0.06Fear 62.9 (15.1) 61.9 (14.6) 0.07Anger 95.5 (5.7) 95.8 (8.9) −0.04Surprise 84.1 (20.9) 77.9 (19.6) 0.31Neutral 67.4 (14.2) 74.0 (15.7) −0.44

a Cohen's d is based on the average standard deviation from the two means.

5K.A. Allott et al. / Schizophrenia Research xxx (2014) xxx–xxx

labeling across varying emotions may be particularly informativewith respect to specific types of psychopathology. It is also possiblethat deficits in recognizing negative emotion reflect epiphenomena ortrait markers of general psychopathology.

The finding that mislabeling neutral emotion was predictive of tran-sition to a psychotic disorder is consistent with studies showing at-riskindividuals to have more difficulty than controls in interpreting facialneutral emotion (Eack et al., 2010; van Rijn et al., 2011). Furthermore,a community sample of psychosis-prone individuals displayed longerreaction times to neutral words after presentation of emotional stimuli(van't Wout et al., 2004). It is noteworthy that the mean accuracy inidentifying neutral faces in the non-transition group was similar to thehealthy control group from our previous study (93.6% and 91.7%,respectively) (Amminger et al., 2012c), suggesting normal neutral emo-tion recognition in those who do not transition to a psychotic disorder,but a deficit in those who transitioned. As opposed to other emotions,neutral emotion may be more susceptible to being perceived as ambig-uous, and for truly prodromal individuals, there may be a tendency tointerpret ambiguous stimuli with undue significance.

Biological theories of psychosis have accounted for the tendency tomisinterpret benign or ambiguous social cues. In psychosis, increaseddopamine is observed in the mesolimbic pathway, with dopaminebeing a key neurochemical determinant of the significance of environ-mental cues to human motivations (Kapur, 2003). Abnormal increasesin this neurotransmitter are proposed to lead to “aberrant salience” ofenvironmental cues (Kapur, 2003). Thus, increased dopamine synthesisin prodromal individuals may influence the perceived salience of envi-ronmental cues, and in this instance, a tendency to interpret neutralfaces as emotionally meaningful. Concordantly, recent findings showelevated dopamine synthesis in prodromal individuals, which is corre-lated with severity of psychotic (but not anxiety or depressive) symp-toms (Howes et al., 2009) and predicts later psychotic disordertransition (Howes et al., 2011).

In the current study, UHR individualswho transitioned to a psychoticdisorder were significantly more likely to label neutral emotion as fearthan those who did not transition. Interestingly, van Rijn et al. (2011)found that UHR participants were more likely than controls to attributeneutral emotion as angry. Patients with schizophrenia have also beenshown to display deficits in neutral emotion recognition, with a greater

Please cite this article as: Allott, K.A., et al., Emotion recognition as a predicSchizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.01.037

tendency than healthy controls to label neutral emotion as disgust andfear (Kohler et al., 2003). Collectively, these findings may suggest an in-creased propensity for interpreting neutral emotion as threatening andsupports further investigation into whether misattribution of neutralemotion is implicated in the development of psychotic disorder. Indeed,psychological models include the role of negative cognitive appraisalsof ambiguous or anomalous perceptual experiences (in the contextof emotional disturbance) in the formation and persistence of positivepsychotic symptoms (Green and Phillips, 2004; Garety et al., 2007).

The result showing more accurate identification of fear in UHRindividuals who transitioned was unexpected. Interestingly, the meanaccuracy in identifying fear in faces in the transition group was similarto the accuracy of the healthy control group from our previous study(81.8% and 81.9%, respectively) (Amminger et al., 2012c), suggestiveof no fear recognition deficit in those who transitioned. A possibleexplanation for this finding is that truly prodromal individuals withevolving positive symptoms may be more attuned to identifying fearthan non-transition participants (Green and Phillips, 2004). Indeed,paranoid schizophrenia patients have been found to be better at identi-fying negative emotions than non-paranoid patients (Kline et al., 1992;Davis and Gibson, 2000). IQ may be another factor accounting for thebetter fear identification in the transition group, given that the meanIQ of the transition group was higher than the non-transition group(albeit not statistically significantly so). Exploration of whether therewas a direct correlation between IQ and fear emotion recognition re-vealed no relationship between these variables (r = − .061, p = .72).Given the relative novelty of the finding of normal fear identificationpredicting transition, replication in other UHR samples including longi-tudinal evaluation of emotion recognition and symptom developmentwill be informative.

Apart from being a potential early clinical marker of psychoticdisorders, emotion recognition abilities may provide insight into theillness' neural basis. Alignedwith our behavioral findings regarding neu-tral emotion recognition, functional imaging studies in schizophreniahave shown increased activation of the hippocampus, parahippocampusand amygdala when viewing neutral faces (Holt et al., 2006; Surguladzeet al., 2006; Hall et al., 2008), providing neural evidence that neutralfaces are perceived as salient by people with schizophrenia. Similarly,Seiferth et al. (2008) found that UHR participants exhibited increased ac-tivation,whereas healthy controls displayed decreased activation to neu-tral faces (contrasted with emotional faces) in classic face-processingregions including the frontal gyri, thalamus, hippocampus and cuneus.These abnormalities were observed despite similar task performanceacross the groups. This suggests that brain functioning underlying emo-tion (particularly neutral) recognition may already be altered before theonset of psychotic disorder and may precede overt behavioral deficits.

The issue as to whether emotion recognition deficits represent spe-cific trait or statemarkers of psychotic disorder cannot be speculated onbased on the current findings. The premise that emotion recognitiondeficits reflect trait markers of psychotic disorder was not completelysupported by the current findings, as better fear identificationwas asso-ciated with transition to a psychotic disorder. It is entirely possible thatthe findings of the current study reflect state-related emotion recogni-tion abilities in association with development of acute psychotic symp-toms. Repeated longitudinal assessment of emotion recognition abilitiesbefore and after the onset of psychotic disorder, including duringpsychotic episode and remission phases is needed to further addressthe question as to whether altered emotion recognition abilities arestate or trait markers of psychotic disorder.

The strengths of this study include the longitudinal designwith a rel-atively complete cohort over the 12-month follow-up period (N90%)(Amminger et al., 2010), participants were antipsychotic-naïve, andemotion recognition was examined across two modalities. However,there are also limitations. First, transition to psychotic disorder wasonly examined over 12 months and psychosis has been shown todevelop in some UHR individuals after this period (Cannon et al.,

tor of transition to a psychotic disorder in ultra-high risk participants,

6 K.A. Allott et al. / Schizophrenia Research xxx (2014) xxx–xxx

2008). Thus, the findings only relate to relatively short-term predictionof psychotic disorder transition. Second, the sample size was relativelysmall and despite the significant findings, results require replication inlarger samples to determine generalizability. Third, social functioningwas not measured, which has been shown to be predictive of transition(Cornblatt et al., 2012) and is associated with emotion recognition inschizophrenia (Irani et al., 2012). Finally, although there was no signif-icant difference in IQ between the two groups, specific neurocognitivedomains that have been shown to be predictive of psychotic disordertransition (Brewer et al., 2006) were not examined.

In conclusion, the current findings show that alterations in emotionrecognition present before the onset of psychosis are associated withlater transition to a psychotic disorder. Specifically, poorer identificationof neutral and better identification of fear in faces was significantly pre-dictive of psychotic disorder transition. Thesefindingsmay indicate thatemotion recognition represents a trait marker of psychotic disorder,lending support to the neurodevelopmental hypothesis. However, fur-ther work is needed to determine when emotion recognition deficitsdevelop, as they may also reflect state-related changes close to theonset of disorder (McGorry et al., 2006). In combination with otherpredictive factors, emotion recognition may enhance the accuracy ofpredicting transition to a psychotic disorder. Preventive interventionstargeting emotion recognition deficits may be indicated in psychosis-prone individuals.

Role of funding sourceThis research was supported by grant # 03T-315 from the Stanley Medical Research

Institute and National Bank of Austria grant # 9848. G.P.A. was supported by grant# 566529 from the National Health and Medical Research Council (NHMRC), Australia;K.A., C.B. and S.B. were supported by postdoctoral clinical research training fellowshipsfrom the NHMRC, Australia (#628884, #567042 and #1036425, respectively); and B.N.was supported by a Career Development Fellowship from the NHMRC, Australia(#1027532).

ContributorsK.A. co-designed the study, undertook the literature search, conducted most of the

statistical analyses and wrote the first draft of the manuscript. G.P.A. co-designed thestudy, wrote the protocol, assisted with the statistical analyses and contributed to thefinal draft of the manuscript. M.R.S. and M.S. collected the data for the study and contrib-uted to the final draft of themanuscript. A.T., B.N., S.B., C.B., P.Mc. and A.B. provided expertadvice on the findings of the study and contributed to the final draft of the manuscript.H.-P.Y. assistedwith the statistical analyses. All authors contributed to and have approvedthe final manuscript.

Conflict of interestAll authors declare they have no financial or other potential conflicts of interest.

AcknowledgmentsNone.

References

Addington, J., Penn, D., Woods, S.W., Addington, D., Perkins, D.O., 2008. Facial affectrecognition in individuals at clinical high risk for psychosis. Br. J. Psychiatry 192(1), 67–68.

Addington, J., Piskulic, D., Perkins, D., Woods, S.W., Liu, L., Penn, D.L., 2012. Affect recogni-tion in people at clinical high risk of psychosis. Schizophr. Res. 140 (1–3), 87–92.

Amminger, G.P., Schafer, M.R., Papageorgiou, K., Klier, C.M., Cotton, S.M., Harrigan, S.M.,Mackinnon, A., McGorry, P.D., Berger, G.E., 2010. Long-chain omega-3 fatty acids forindicated prevention of psychotic disorders: a randomized, placebo-controlled trial.Arch. Gen. Psychiatry 67 (2), 146–154.

Amminger, G.P., Schäfer, M.R., Klier, C.M., Schlögelhofer, M., Mossaheb, N., Thompson, A.,Bechdolf, A., Allott, K., McGorry, P.D., Nelson, B., 2012a. Facial and vocal affect percep-tion in people at ultra-high risk of psychosis, first-episode schizophrenia and healthycontrols. Early Interv. Psychiatry 6 (4), 450–454.

Amminger, G.P., Schafer, M.R., Klier, C.M., Slavik, J.M., Holzer, I., Holub, M., Goldstone, S.,Whitford, T.J., McGorry, P.D., Berk, M., 2012b. Decreased nervonic acid levels inerythrocyte membranes predict psychosis in help-seeking ultra-high-risk individuals.Mol. Psychiatry 17 (12), 1150–1152.

Amminger, G.P., Schafer, M.R., Papageorgiou, K., Klier, C.M., Schlogelhofer, M., Mossaheb,N., Werneck-Rohrer, S., Nelson, B., McGorry, P.D., 2012c. Emotion recognition inindividuals at clinical high-risk for schizophrenia. Schizophr. Bull. 38 (5),1030–1039.

Please cite this article as: Allott, K.A., et al., Emotion recognition as a predicSchizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.01.037

Andreasen, N.C., Endicott, J., Spitzer, R.L., Winokur, G., 1977. Family history methodusing diagnostic criteria — reliability and validity. Arch. Gen. Psychiatry 34 (10),1229–1235.

Bediou, B., Asri, F., Brunelin, J., Krolak-Salmon, P., D'Amato, T., Saoud, M., Tazi, I., 2007.Emotion recognition and genetic vulnerability to schizophrenia. Br. J. Psychiatry191, 126–130.

Brewer, W.J., Wood, S.J., Phillips, L.J., Francey, S.M., Pantelis, C., Yung, A.R., Cornblatt, B.,McGorry, P.D., 2006. Generalized and specific cognitive performance in clinicalhigh-risk cohorts: a review highlighting potential vulnerabilitymarkers for psychosis.Schizophr. Bull. 32 (3), 538–555.

Cannon, T.D., Cadenhead, K., Cornblatt, B., Woods, S.W., Addington, J., Walker, E., Seidman,L.J., Perkins, D., Tsuang, M., McGlashan, T., Heinssen, R., 2008. Prediction of psychosisin youth at high clinical risk: a multisite longitudinal study in North America. Arch.Gen. Psychiatry 65 (1), 28–37.

Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences, 2nd ed. LawrenceEarlbaum Associates, Hillsdale, NJ.

Cornblatt, B.A., Carrion, R.E., Addington, J., Seidman, L., Walker, E.F., Cannon, T.D.,Cadenhead, K.S., McGlashan, T.H., Perkins, D.O., Tsuang, M.T., Woods, S.W.,Heinssen, R., Lencz, T., 2012. Risk factors for psychosis: impaired social and rolefunctioning. Schizophr. Bull. 38 (6), 1247–1257.

Couture, S.M., Penn, D.L., Roberts, D.L., 2006. The functional significance of social cognitionin schizophrenia: a review. Schizophr. Bull. 32 (S1), S44–S63.

Davis, P.J., Gibson, M.G., 2000. Recognition of posed and genuine facial expressions ofemotion in paranoid and nonparanoid schizophrenia. J. Abnorm. Psychol. 109 (3),445–450.

Demenescu, L.R., Kortekaas, R., den Boer, J.A., Aleman, A., 2010. Impaired attribution ofemotion to facial expressions in anxiety and major depression. PLoS One 5 (12),e15058.

Derntl, B., Seidel, E.M., Kryspin-Exner, I., Hasmann, A., Dobmeier, M., 2009. Facial emotionrecognition in patients with bipolar I and bipolar II disorder. Br. J. Clin. Psychol. 48,363–375.

Eack, S.M., Mermon, D.E., Montrose, D.M., Miewald, J., Gur, R.E., Gur, R.C., Sweeney, J.A.,Keshavan, M.S., 2010. Social cognition deficits among individuals at familial highrisk for schizophrenia. Schizophr. Bull. 36 (6), 1081–1088.

Edwards, J., Pattison, P.E., Jackson, H.J., Wales, R.J., 2001. Facial affect and affective prosodyrecognition in first-episode schizophrenia. Schizophr. Res. 48, 235–253.

Edwards, J., Jackson, H.J., Pattison, P.E., 2002. Emotion recognition via facial expressionand affective prosody in schizophrenia: a methodological review. Clin. Psychol. Rev.22, 789–832.

Feinberg, T.E., Rifkin, A., Schaffer, C., Walker, E., 1986. Facial discrimination and emotionalrecognition in schizophrenia and affective disorders. Arch. Gen. Psychiatry 43 (3),276–279.

First, M.B., Spitzer, R.L., Gibbon,M., Williams, J.B.W., 2001. Structured Clinical Interview forDSM-IV-TR Axis I Disorders — Patient Edition (SCID-I/P, 2/2001 Revision) BiometricsResearch Department. New York State Psychiatric Institute, New York.

Garety, P.A., Bebbington, P., Fowler, D., Freeman, D., Kuipers, E., 2007. Implications forneurobiological research of cognitive models of psychosis: a theoretical paper.Psychol. Med. 37 (10), 1377–1391.

Green, M.J., Phillips, M.L., 2004. Social threat perception and the evolution of paranoia.Neurosci. Biobehav. Rev. 28 (3), 333–342.

Green, M.F., Bearden, C.E., Cannon, T.D., Fiske, A.P., Hellemann, G.S., Horan, W.P., Kee, K.,Kern, R.S., Lee, J., Sergi, M.J., Subotnik, K.L., Sugar, C.A., Ventura, J., Yee, C.M.,Nuechterlein, K.H., 2012. Social cognition in schizophrenia, part 1: performanceacross phase of illness. Schizophr. Bull. 38 (4), 854–864.

Hall, J., Whalley, H.C., McKirdy, J.W., Romaniuk, L., McGonigle, D., McIntosh, A.M., Baig, B.J.,Gountouna, V.-E., Job, D.E., Donaldson, D.I., Sprengelmeyer, R., Young, A.W.,Johnstone, E.C., Lawrie, S.M., 2008. Overactivation of fear systems to neutral faces inschizophrenia. Biol. Psychiatry 64 (1), 70–73.

Holt, D.J., Kunkel, L., Weiss, A.P., Goff, D.C., Wright, C.I., Shin, L.M., Rauch, S.L., Hootnick, J.,Heckers, S., 2006. Increased medial temporal lobe activation during the passiveviewing of emotional and neutral facial expressions in schizophrenia. Schizophr.Res. 82 (2–3), 153–162.

Howes, O.D., Montgomery, A.J., Asselin, M.C., Murray, R.M., Valli, I., Tabraham, P., Bramon-Bosch, E., Valmaggia, L., Johns, L., Broome, M., McGuire, P.K., Grasby, P.M., 2009.Elevated striatal dopamine function linked to prodromal signs of schizophrenia.Arch. Gen. Psychiatry 66 (1), 13–20.

Howes, O.D., Bose, S.K., Turkheimer, F., Valli, I., Egerton, A., Valmaggia, L.R., Murray, R.M.,McGuire, P., 2011. Dopamine synthesis capacity before onset of psychosis: a prospec-tive (18)F-DOPA PET imaging study. Am. J. Psychiatry 168 (12), 1311–1317.

Irani, F., Seligman, S., Kamath, V., Kohler, C., Gur, R.C., 2012. A meta-analysis of emotionperception and functional outcomes in schizophrenia. Schizophr. Res. 137 (1–3),203–211.

Kapur, S., 2003. Psychosis as a state of aberrant salience: a framework linking biology, phe-nomenology, and pharmacology in schizophrenia. Am. J. Psychiatry 160 (1), 13–23.

Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale(PANSS) for schizophrenia. Schizophr. Bull. 13 (2), 261–276.

Kee, K.S., Horan, W.P., Mintz, J., Green, M.F., 2004. Do the siblings of schizophreniapatients demonstrate affect perception deficits? Schizophr. Res. 67 (1), 87–94.

Kline, J.S., Smith, J.E., Ellis, H.C., 1992. Paranoid and nonparanoid schizophrenic processingof facially displayed affect. J. Psychiatr. Res. 26 (3), 169–182.

Kohler, C.G., Turner, T.H., Bilker, W.B., Brensinger, C.M., Siegel, S.J., Kanes, S.J., Gur, R.E.,Gur, R.C., 2003. Facial emotion recognition in schizophrenia: intensity effects anderror pattern. Am. J. Psychiatry 160 (10), 1768–1774.

Kohler, C.G., Walker, J.B., Martin, E.A., Healey, K.M., Moberg, P.J., 2010. Facial emotionperception in schizophrenia: a meta-analytic review. Schizophr. Bull. 36 (5),1009–1019.

tor of transition to a psychotic disorder in ultra-high risk participants,

7K.A. Allott et al. / Schizophrenia Research xxx (2014) xxx–xxx

Kohler, C.G., Hoffman, L.J., Eastman, L.B., Healey, K., Moberg, P.J., 2011. Facial emotionperception in depression and bipolar disorder: a quantitative review. PsychiatryRes. 188 (3), 303–309.

Kucharska-Pietura, K., David, A.S., Masiak, M., Phillips, M.L., 2005. Perception of facial andvocal affect by people with schizophrenia in early and late stages of illness. Br.J. Psychiatry 187, 523–528.

Marenco, S.,Weinberger, D.R., 2000. The neurodevelopmental hypothesis of schizophrenia:following a trail of evidence from cradle to grave. Dev. Psychopathol. 12 (3), 501–527.

McGorry, P.D., Hickie, I.B., Yung, A.R., Pantelis, C., Jackson, H.J., 2006. Clinical staging ofpsychiatric disorders: a heuristic framework for choosing earlier, safer and moreeffective interventions. Aust. N. Z. J. Psychiatry 40, 616–622.

Montgomery, S.A., Asberg, M., 1979. New depression scale designed to be sensitive tochange. Br. J. Psychiatry 134 (APR), 382–389.

Morrison, A.P., French, P., Walford, L., Lewis, S.W., Kilcommons, A., Green, J., Parker, S.,Bentall, R.P., 2004. Cognitive therapy for the prevention of psychosis in people atultra-high risk — randomised controlled trial. Br. J. Psychiatry 185, 291–297.

Oswald, W.D., Roth, E., 1987. Number Combination Test (Zahlen-Verbindungs-Test; ZVT).Testzentrale, Gottingen, Germany.

Pinkham, A.E., Penn, D.L., Perkins, D.O., Graham, K.A., Siege, M., 2007. Emotion perceptionand social skill over the course of psychosis: a comparison of individuals “at-risk” forpsychosis and individuals with early and chronic schizophrenia spectrum illness.Cogn. Neuropsychiatry 12, 198–212.

Please cite this article as: Allott, K.A., et al., Emotion recognition as a predicSchizophr. Res. (2014), http://dx.doi.org/10.1016/j.schres.2014.01.037

Seiferth, N.Y., Pauly, K., Habel, U., Kellermann, T., Shah, N.J., Ruhrmann, S., Klosterkotter, J.,Schneider, F., Kircher, T., 2008. Increased neural response related to neutral faces inindividuals at risk for psychosis. NeuroImage 40 (1), 289–297.

Surguladze, S., Russell, T., Kucharska-Pietura, K., Travis, M.J., Giampietro, V., David, A.S.,Phillips, M.L., 2006. A reversal of the normal pattern of parahippocampal responseto neutral and fearful faces is associated with reality distortion in schizophrenia.Biol. Psychiatry 60 (5), 423–431.

Thompson, A.D., Bartholomeusz, C., Yung, A.R., 2011. Social cognition deficits and the‘ultra high risk’ for psychosis population: a review of literature. Early Interv. Psychiatry5 (3), 192–202.

van Rijn, S., Aleman, A., de Sonneville, L., Sprong, M., Ziermans, T., Schothorst, P., vanEngeland, H., Swaab, H., 2011. Misattribution of facial expressions of emotion inadolescents at increased risk of psychosis: the role of inhibitory control. Psychol.Med. 41 (3), 499–508.

Van't Wout, M., Aleman, A., Kessels, R.P.C., Laroi, F., Kahn, R.S., 2004. Emotional processingin a non-clinical psychosis-prone sample. Schizophr. Res. 68 (2–3), 271–281.

Yung, A.R., Phillips, L.J., McGorry, P.D., McFarlane, C.A., Francey, S., Harrigan, S., Patton,G.C., Jackson, H.J., 1998. Prediction of psychosis — a step towards indicated preven-tion of schizophrenia. Br. J. Psychiatry 172, 14–20.

Yung, A.R., Yuen, H.-P., Berger, G.E., Francey, S., Hung, T., Phillips, L.J., McGorry, P.D., 2007.Declining transition rate in Ultra High Risk (prodromal) services: dilution or reduc-tion of risk? Schizophr. Bull. 33, 673–681.

tor of transition to a psychotic disorder in ultra-high risk participants,